Research map / Video / Video segmentation and tracking
Recurrent Convolutional Neural Networks for Continuous Sign Language Recognition by Staged Optimization
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Add the alignment-proposal feature-tuning stage and detection-net regularization (ablation) | Our-end2end · RWTH-PHOENIX-Weather 2014 test | WER: 46.9 -> 38.7 | Helped |
| Add the detection net to BLSTM sequence learning (ablation) | BLSTM with tuned features · RWTH-PHOENIX-Weather 2014 test | WER: 41.9 -> 38.7 | Helped |
| Replace 3D-CNN feature extraction with temporal convolution and pooling (ablation) | BLSTM end-to-end model · RWTH-PHOENIX-Weather 2014 test | WER: 77.6 -> 47.3 | Helped |
| Add PN-Net pretraining on right-hand patches (ablation) | ConvTC+BLSTM · RWTH-PHOENIX-Weather 2014 test | WER: 47.3 -> 46.9 | Helped |
| Use BLSTM rather than RNN (ablation) | ConvTC sequence model · RWTH-PHOENIX-Weather 2014 test | WER: 53.7 -> 47.3 | Helped |
| Use BLSTM rather than LSTM (ablation) | ConvTC sequence model · RWTH-PHOENIX-Weather 2014 test | WER: 51.3 -> 47.3 | Helped |
| Use BLSTM rather than RNN (ablation) | Sequence model with tuned feature representations · RWTH-PHOENIX-Weather 2014 test | WER: 44.8 -> 41.9 | Helped |
| Use BLSTM rather than LSTM (ablation) | Sequence model with tuned feature representations · RWTH-PHOENIX-Weather 2014 test | WER: 43.6 -> 41.9 | Helped |
| Use the proposed model without extra supervision or imported initial alignment | Continuous sign recognition on RWTH-PHOENIX-Weather 2014 · RWTH-PHOENIX-Weather 2014 test | WER: 38.8 vs 38.7 | Mixed |
Closest papers by meaning
- Iterative Alignment Network for Continuous Sign Language Recognition (CVPR 2019 · similarity 0.92)
- TCNet: Continuous Sign Language Recognition from Trajectories and Correlated Regions (AAAI 2024 · similarity 0.90)
- Bridging Sign and Spoken Languages: Pseudo Gloss Generation for Sign Language Translation (NeurIPS 2025 · similarity 0.87)
- Improving Gloss-free Sign Language Translation by Reducing Representation Density (NeurIPS 2024 · similarity 0.86)
- Cross-Sentence Gloss Consistency for Continuous Sign Language Recognition (AAAI 2024 · similarity 0.86)
- KD-MSLRT: Lightweight Sign Language Recognition Model Based on Mediapipe and 3D to 1D Knowledge Distillation (AAAI 2025 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.